The fibroblast growth factor (FGF) family is characterized by 22 genetically distinct, homologous ligands, which are grouped into seven subfamilies. FGF-21 is most closely related to, and forms a subfamily with, FGF-19 and FGF-23. This FGF subfamily regulates diverse physiological processes uncommon to classical FGFs, namely energy and bile acid homeostasis, glucose and lipid metabolism, and phosphate as well as vitamin D homeostasis. Moreover, unlike other FGFs, this subfamily acts in an endocrine fashion (Moore, D. D. (2007) Science 316, 1436-8) (Beenken et al. (2009) Nature Reviews Drug Discovery 8, 235).
FGF21 is a 209 amino acid polypeptide containing a 28 amino acid leader sequence (SEQ ID NO:132). Human FGF21 has about 79% amino acid identity to mouse FGF21 and about 80% amino acid identity to rat FGF21. Fibroblast growth factor 21 (FGF21) has been described as a treatment for ischemic vascular disease, wound healing, and diseases associated with loss of pulmonary, bronchia or alveolar cell function (Nishimura et al. (2000) Biochimica et Biophysica Acta, 1492:203-206; patent publication WO01/36640; and patent publication WO01/18172). Although FGF-21 activates FGF receptors and downstream signaling molecules, including FRS2a and ERK, direct interaction of FGFRs and FGF-21 has not been detected. Studies have identified β-klotho, which is highly expressed in liver, adipocytes and pancreas, as a determinant of the cellular response to FGF-21 and a cofactor which mediates FGF-21 signaling through FGFRs (Kurosu, H. et al. (2007) J Biol Chem 282, 26687-95). FGF21 is a potent agonist of the FGFR1(IIIc), FGFR2(IIIc) and FGFR3(IIIc) β-klotho signaling complexes.
FGF-21 has been shown to induce insulin-independent glucose uptake. FGF-21 has also been shown to ameliorate hyperglycemia in a range of diabetic rodent models. In addition, transgenic mice over-expressing FGF-21 were found to be resistant to diet-induced metabolic abnormalities, and demonstrated decreased body weight and fat mass, and enhancements in insulin sensitivity (Badman, M. K. et al. (2007) Cell Metab 5, 426-37). Administration of FGF-21 to diabetic non-human primates caused a decline in fasting plasma glucose, triglycerides, insulin and glucagon levels, and led to significant improvements in lipoprotein profiles, including a nearly 80% increase in HDL cholesterol (Kharitonenkov, A. et al. (2007) Endocrinology 148, 774-81). Recent studies investigating the molecular mechanisms of FGF21 action have identified FGF21 as an important endocrine hormone that helps to control adaptation to the fasting state (Badman et al. (2009) Endocrinology 150, 4931)(Inagaki et al. (2007) Cell Metabolism 5, 415). This provides a previously missing link downstream of PPARα, by which the liver communicates with the rest of the body in regulating the biology of energy homeostasis (Galman et al. (2008) Cell Metabolism 8, 169) (Lundasen et al. (2007) Biochemical and Biophysical Research Communications 360, 437).
FGF21 regulates adipocyte homeostasis through activation of an AMPK/SIRT1/PGC1α pathway to inhibit PPARγ expression and increase mitochondrial function (Chau et al. (2010) PNAS 107, 12553). FGF21 also increases glucose uptake by skeletal muscle as measured in cultured human myotubes and isolated mouse tissue (Mashili et al. (2011) Diabetes Metab Res Rev 27, 286-97). FGF21 treatment of rodent islet cells leads to improved function and survival through activation of ERK1/2 and Akt pathways (Wente et al. (2006) Diabetes 55, 2470). FGF21 treatment also results in altered gene expression for lipogenesis and fatty acid oxidation enzymes in rodent livers, likely through HNF4α and Foxa2 signaling. However, recent studies (Wei et al. (2012) PNAS 109, 3143-48) indicate that treatment of diet-induced obese mice with FGF21 induces bone loss, due to a diminished inactivation of PPARγ (via reduced sumoylation); a shift of mesenchymal stem cell differentiation from osteoblasts to adipocytes is seen in the presence of increased PPARγ activity in the bone following FGF21 treatment.
A difficulty associated with using FGF-21 directly as a biotherapeutic is that its half-life is very short (Kharitonenkov, A. et al. (2005) Journal of Clinical Investigation 115:1627-1635). In mice, the half-life of human FGF21 is 0.5 to 1 hours, and in cynomolgus monkeys, the half-life is 2 to 3 hours. FGF21 may be utilized as a multi-use, sterile pharmaceutical formulation. However, it has been determined that preservatives, e.g., m-cresol, have an adverse effect on its stability under these conditions.
Another potent metabolic regulator already represented in the clinic is Glucagon-Like Peptide-1 (GLP-1) (Knudsen et al. (2004) Journal of Medicinal Chemistry 47, 4128). GLP-1 is a 36 amino acid incretin secreted by L-cells of the mammalian gut, acting on both alpha and beta cells to stimulate insulin secretion and inhibit glucagon release in a glucose-dependent manner (Hare et al. (2010) Diabetes 59, 1765; Meier et al. (2005) Diabetes-Metabolism Research and Reviews 21, 91). GLP-1 binds to and activates the GLP-1 receptor (GLP-1R), a seven-transmembrane helix protein of the class II family of G-protein coupled receptors (GPCRs) (Mayo et al. (2003) Pharmacological Reviews 55:167). As a GLP-1 receptor agonist, GLP-1 has an important role in decreasing post-prandial blood glucose levels by stimulating insulin secretion from the pancreas in order to increase glucose absorption in the peripheral tissues and inhibiting glucagon secretion, resulting in reduced hepatic glucose release.
A second clinically important GLP-1 receptor agonist is Exendin-4. Exendin-4 is a 39 residue polypeptide produced in the salivary glands of the Gila Monster lizard (Goke et al. (1993) Diabetes 46:433-439; Fehmann et al. (1995) Endocrine Rev. 16:390-410). Although it is the product of a uniquely non-mammalian gene and appears to be expressed only in the salivary gland, Exendin-4 shares a 52% amino acid sequence homology with GLP-1, and in mammals interacts with the GLP-1 receptor (Goke, et al.; Thorens et al. (1993) Diabetes 42:1678-1682). In vitro, Exendin-4 has been shown to promote insulin secretion by insulin producing cells and, given in equimolar quantities, is more potent than GLP-1 at causing insulin release from insulin producing cells. Furthermore, Exendin-4 potently stimulates insulin release to reduce plasma glucose levels in both rodents and humans and is longer acting than GLP-1; however, because it does not occur naturally in mammals, Exendin-4 has certain potential antigenic properties in mammals that GLP-1 lacks.
The ability of GLP-1 and Exendin-4 analogues (e.g., Liraglutide and Byetta) to improve glucose control in humans is established in the clinic (Idris (2010) Diabetes Obesity & Metabolism 12, 89; Monami et al (2009) European Journal of Endocrinology 160, 909). GLP-1 has also been reported to increase beta cell mass both through induced proliferation and inhibition of apoptosis (Egan, A et al (2003) Diabetes-Metabolism Research and Reviews 19, 115; Farilla, L. et al. (2003) Endocrinology 144, 5149; Xu, G. et al. (1999) Diabetes 48, 2270). It also acts as an intestinal hormone to inhibit acid secretion and gastric emptying in the stomach while providing a satiety signal that decreases appetite (Vilsboll et al. (2009) Best Practice & Research Clinical Endocrinology & Metabolism 23, 453). These effects likely account for beneficial weight loss observed with administration of GLP-1 analogues to type 2 diabetes patients. GLP-1 has also been shown to be cardioprotective in postischemic rodent hearts (Ossum et al. (2009) Pharmacological Research 60, 411; Sonne, D. P. et al. (2008) Regulatory Peptides 146, 243; Nikolaidis, L. A. et al. (2004) Circulation 109, 962).
Additionally, GLP-1 can reduce the differentiation of human mesenchymal stem cells (hMSCs) to adipocytes by reducing the expression of PPARγ, and GLP-1 promotes cellular proliferation and cytoprotection of hMSCs (Sanz et al. (2010) Am J Physiol Endocrinol Metab 298, E634-E643).
In developing an FGF21 protein, including a variant or analogue thereof, for use as a therapeutic in the treatment of type 1 and type 2 diabetes mellitus and other metabolic conditions, an increase in half-life and stability would be desirable. FGF21 proteins having enhanced half-life and stability would allow for less frequent dosing of patients being administered the protein. Clearly, there is a need to develop a stable aqueous protein formulation for the therapeutic protein FGF21.
Furthermore, a significant challenge in the development of protein pharmaceuticals, such as metabolic regulators FGF21, GLP-1, and Exendin-4, is to cope with their physical and chemical instabilities. The compositional variety and characteristics of proteins define specific behaviors such as folding, conformational stability, and unfolding/denaturation. Such characteristics should be addressed when aiming to stabilize proteins in the course of developing pharmaceutical formulation conditions utilizing aqueous protein solutions (Wang, W., Int. J. of Pharmaceutics, 18, (1999)). A desired effect of stabilizing therapeutic proteins of interest, e.g., the proteins of the present invention, is increasing resistance to proteolysis and enzymatic degradation, thereby improving protein stability and reducing protein aggregation.